init inference

app.py

@@ -1,53 +1,55 @@
 import gradio as gr
-from transformers import AutoModelForCausalLM, AutoTokenizer
-
-# Load your model and tokenizer
-model_name = "Mat17892/llama_lora_G14"  # Replace with your Hugging Face model name
-tokenizer = AutoTokenizer.from_pretrained(model_name)
-model = AutoModelForCausalLM.from_pretrained(model_name)
-
-def respond(
-    message,
-    history: list[tuple[str, str]],
-    system_message,
-    max_tokens,
-    temperature,
-    top_p,
-):
-    messages = [{"role": "system", "content": system_message}]
-
-    for val in history:
-        if val[0]:
-            messages.append({"role": "user", "content": val[0]})
-        if val[1]:
-            messages.append({"role": "assistant", "content": val[1]})
-
-    messages.append({"role": "user", "content": message})
-
-    # Prepare input for the model
-    input_text = message
-    inputs = tokenizer(input_text, return_tensors="pt")
-    
-    # Generate response
-    outputs = model.generate(
-        **inputs, 
-        max_new_tokens=max_tokens, 
-        temperature=temperature, 
-        top_p=top_p
+from llama_cpp import Llama
+
+# Path to the GGUF model file
+model_path = "Mat17892/lora_llama_gguf_g14/model.gguf"  # Update this path to your model
+
+# Load the GGUF model using llama-cpp-python
+print("Loading model...")
+llm = Llama(model_path=model_path, n_ctx=2048, n_threads=8)  # Adjust threads as needed
+print("Model loaded!")
+
+# Chat function
+def chat_with_model(user_input, chat_history):
+    """
+    Process user input and generate a response from the model.
+    :param user_input: User's input string
+    :param chat_history: Conversation history
+    :return: Updated chat history
+    """
+    # Format chat history for the Llama model
+    prompt = ""
+    for turn in chat_history:
+        prompt += f"User: {turn['user']}\nAI: {turn['ai']}\n"
+    prompt += f"User: {user_input}\nAI:"
+
+    # Generate response from the model
+    response = llm(prompt)["choices"][0]["text"].strip()
+
+    # Update chat history
+    chat_history.append({"user": user_input, "ai": response})
+    return chat_history, chat_history
+
+# Gradio UI
+with gr.Blocks() as demo:
+    gr.Markdown("# 🦙 LLaMA GGUF Chatbot")
+    chat_box = gr.Chatbot(label="Chat with the GGUF Model")
+
+    with gr.Row():
+        with gr.Column(scale=4):
+            user_input = gr.Textbox(label="Your Message", placeholder="Type a message...")
+        with gr.Column(scale=1):
+            submit_btn = gr.Button("Send")
+
+    chat_history = gr.State([])
+
+    # Link components
+    submit_btn.click(
+        chat_with_model,
+        inputs=[user_input, chat_history],
+        outputs=[chat_box, chat_history],
+        show_progress=True,
     )
-    response = tokenizer.decode(outputs[0], skip_special_tokens=True)
-    return response
-
-# Create the Gradio interface
-demo = gr.ChatInterface(
-    respond,
-    additional_inputs=[
-        gr.Textbox(value="You are a friendly chatbot.", label="System message"),
-        gr.Slider(minimum=1, maximum=2048, value=512, step=1, label="Max new tokens"),
-        gr.Slider(minimum=0.1, maximum=4.0, value=0.7, step=0.1, label="Temperature"),
-        gr.Slider(minimum=0.1, maximum=1.0, value=0.95, step=0.05, label="Top-p"),
-    ],
-)
-
-if __name__ == "__main__":
-    demo.launch()
+
+# Launch the app
+demo.launch()

requirements.txt

@@ -1,4 +1,3 @@
 huggingface_hub==0.25.2
 gradio 
-transformers
-torch
+llama-cpp-python